|
Class Summary |
| CondEntropyIndicatorSelector |
The conditonal entropy indicator evaluator. |
| MaxEntIterativeIndicatorSel |
MaxEntIterativeIndicatorSel selects the highest weigthed feature (by MaxEnt trainng) for the positive class in each iteration and tries to iteratievly cover the positive class. |
| MaxEntSelector |
MaxEntSelector trains a MaxEnt on the dataset and returns with the features recieving the highest weights for the positive class during the training. |
| MutualInfoIndicatorSelector |
The mutual information indicator selector. |
| ProbIndSel |
Simply the conditional probability of the positive class p(+|w) |
| TresholdOptimizer |
Each indicator selection method has a paramter T. |